Some general question about processing sequential data

[LauraCraft]

I just finished the course about [Sequences, Time Series and Prediction] and I have a few questions:

• Isn’t the attention model better for sequence data than an RNN / LSTM? When is it recommended to use the attention model and when is an RNN / LSTM recommended?

• in the course we only saw examples of calculating the next day’s value by yesterday’s value, what happens when there are more parameters that affect the next day value that I would like to account for?

[Deepti_Prasad]

@LauraCraft

attention model

even attention model use rnn/lstm layer to make sure the sequential data is correlated with query, key, value.

so the

@LauraCraft

the course we only saw examples of calculating the next day’s value by yesterday’s value, what happens when there are more parameters that affect the next day value that I would like to account for

if this is related features, then input for sequential data would be based on parameters involved.